Safia Al-Marhoobi

Research

Thesis

Time series analysis and forecasting with applications in climate science

The aim of the project is to investigate different approaches of analysing and forecasting time series with particular range of applications in climate science. The main method to be applied is the so-called Singular Spectrum Analysis which is based on approximation of a real-life series by a series of finite rank satisfying some linear recurrent relation. Study of sensitivity of the forecasting models to small perturbations in the data will be studied and used for assessing accuracy of forecasts. The main data used for testing the methodology will be climatic data gathered in Oman.

Supervisors